OpenAI's $5M Astra Defense Beaten by 3 Guys with $5K of Opus, Argues Viral Thread
harris_edouard · x · 2026-09-19
- a16z quoted Greg Brockman saying OpenAI pulled 25% of its production engineers to red-team its own systems with Astra, claiming all P0 vulnerabilities Astra could find were fixed, with new rounds planned per model.
- @Laneless countered that this defense stack — Astra plus a quarter of OpenAI's production engineers plus a conservative $5M in inference — was defeated by roughly three people with $5K of Claude Opus inference on offense, concluding everything might be cooked.
- The debate highlights the stark offense-defense cost asymmetry in AI-era cybersecurity.
Related event: OpenAI Pulled 25% of Engineers to Hunt Bugs with Astra Until P0s Hit Zero(2 posts)→
More from Safety
- WaPo: US AI firms restrict cybersecurity help while Chinese models empower hackers — andreamichi · 2026-09-19
- DepthFirst Labs found TikTok flaws exposing camera, photos and payment data — andreamichi · 2026-09-19
- 'AI lets you get to a bad idea faster': the speed-vs-reliability lesson — paul_scharre · 2026-09-19
- US military had close call after AI produced false intelligence report, CNN reports — paul_scharre · 2026-09-19
- Supply chain backdoors may be impossible to fully purge, researchers urge compiler security — jd_pressman · 2026-09-19
- HEIF Heist: one libheif image-parser bug hit OpenAI, Slack, Meta, GitHub Enterprise and more — anshulkundaje · 2026-09-19